Non-invasive Photoacoustic Imaging of Skin Inflammatory Disorders With Machine Learning-assisted Scoring

Non-invasive Photoacoustic Imaging of Skin Inflammatory Disorders With Machine Learning-assisted Scoring

Non-invasive Photoacoustic Mesoscopic Imaging of Clinical Skin Inflammatory Disorders Coupled With Artificial Intelligence-assisted Scoring to Evaluate Disease Prognosis and Associated Metabolic Comorbidities

Inflammatory skin disorders are usually assessed by disease scoring system such as Scoring AD (SCORAD)/Eczema Area and Severity Index (EASI) and Psoriasis Area and Severity Index (PASI) for atopic eczema and psoriasis respectively. The current approach to score the severity of these inflammatory skin disorders is through clinical observations and questionnaires. These scores however do not reflect the structural characteristics of the skin such as morphology, vasculature architecture and dermis thickness and are subject to inter and intra-assessor variability. Objective inflammatory diseases indicators through non-invasive imaging techniques have the potential to be an important clinical tool to shed light on its severity in an objective manner. Furthermore, given the abundance of cutaneous vasculature, non-invasive imaging in patients with chronic inflammatory skin conditions allows the investigators to evaluate in detail how co-morbidities of metabolic syndrome, especially type 2 diabetes, further affects the vasculature or the epidermis in the skin. It helps to answer the question of whether a tighter control of the "overlying" skin condition helps in management of the underlying co-morbidities. Currently, there are many skin imaging modalities available to visualize the morphology and vascular architecture non-invasively, but they are hindered by their penetration depth and lack of contrast. Examples include optical coherence tomography (OCT), high-frequency ultrasound, and Doppler based ultrasound. In this study, these shortcomings will be circumvented through the usage of photoacoustic mesoscopic imaging, a non-invasive, high resolution, intrinsic or contrast-enhanced imaging technique, which can provide functional and metabolic information at greater depths, and an optical fibre-based handheld confocal Raman spectroscopy system with inbuilt data processing algorithms and software, which allows for highly effective and accurate analysis of various skin constituents, such as ceramides, filaggrin, and hydration. These technologies will allow the investigators to study inflammatory and skin barrier markers in, as well as correlations between, psoriasis, eczema, diabetes, and obesity. In addition, by studying the skin before and after therapeutic interventions, this study will aid in understanding the mechanisms of action and efficacy of various interventions.

Inflammatory skin diseases are increasingly common in various industrialized western societies. In 2015, eczema or dermatitis, the most common inflammatory skin disease, is estimated to have affected 245 million people worldwide. In fact, the biggest percentage of cases seen in National Skin Centre (NSC), Singapore is for dermatitis (34.1%) while other inflammatory skin diseases such as contact dermatitis and psoriasis account for 4.7% and 3.3% of the cases respectively. Inflammatory skin diseases can cause remodeling of the vasculature. This vascular remodeling is brought about by the imbalance between pro- and anti-angiogenic mediators under conditions of chronic inflammation, resulting in either vessel growth or recession. Vascular remodeling of affected skin usually display vascular enlargement during inflammation. In psoriasis, for example, the skin capillaries expand and become tortuous, making the lesions appear red due to the thinned epithelium. Emerging evidence reports that chronic inflammatory skin diseases are closely related to systemic complications such as atherosclerosis, type 2 diabetes or metabolic syndrome due to systemic inflammation. One possible theory linking cutaneous and systemic vascular diseases is from the release of products such as inflammatory cytokines produced in affected skin lesions into the systemic circulation, resulting in the increased risk of inflammation in other organs or tissues. It was reported that individuals with Atopic Dermatitis (AD) in the adult US population tend to have a self-reported history of hypertension and adult-onset diabetes even with control of body mass index and other comorbidities. An analysis of different studies indicated that both individuals from Asia and North America with pediatric and adult-onset of AD have a higher chance of being overweight, which can lead to other metabolic diseases. It was also reported that there is an increased number of inflammatory cells around vessels seen in histopathology biopsies of the human forearm skin for diabetic patients compared to non-diabetic populations. The density of vascular network also tended to be higher in diabetic patients in the forearms compared to non-diabetic subjects. However, no difference in vascular networks were observed between subjects with Type 1 and Type 2 diabetes. The mechanism to explain the increased vessel density in diabetic subjects is not known, though inflammation plays a role due to the presence of inflammatory cells. Since inflammatory skin diseases are systemic disorders due to the co-morbidities, it would shed light on the understanding of these diseases and their linkages to metabolic disease such as diabetes.

All subjects will first have their basic history and clinical measurements taken, followed by skin physiology measurements and photoacoustic mesoscopic imaging as well as confocal Raman spectroscopy measurements.

Quantify various photoacoustic imaging based parameter variations characterising psoriasis, eczema, and diabetes mellitus (total blood volume, epidermis thickness, size of vessels)

The quantified parameters characterizing the skin inflammatory diseases will be aggregated to a novel imaging-based scoring index called novel Eczema Vascular and Structural Index (EVSI, min=0, max= 25, higher scores refer to higher severity of eczema).

Quantify concentration of natural moisturizing factors in the skin via Raman spectroscopy measurements to characterize psoriasis, eczema, and diabetes mellitus

Raman spectroscopy measures the vibrational modes of the chemical molecules in the skin upon laser excitation, resulting in 'fingerprint' Raman-scattered photons. This allows semi-quantitative estimation of the relative concentration of natural moisturizing factors such as urocanic acid, ceramide and water content in the skin.

Novel Eczema Biochemical Index (EBI), based on the biochemical information in skin, formulated using the skin constituents measured from the subjects

The EBI scoring system is used to stage disease severity in these subjects. Min=0, max= 80, higher scores refer to higher severity of eczema

Inclusion Criteria: Patients with a clinical diagnosis of atopic dermatitis based on either the United Kingdom Working Party criteria or the Hanifin and Rajka criteria, age 21 years and above, with or without diabetes mellitus Patients with diagnosis of active chronic plaque psoriasis clinically by managing dermatologist based on typical clinical features of psoriasis, with or without diabetes mellitus Healthy controls of age 21 years and above with no known chronic skin conditions, with or without diabetes mellitus Willingness to participate in the study and undergo skin physiological assessments, photoacoustic mesoscopic imaging, and confocal Raman spectroscopy Exclusion Criteria: Mentally incompetent, younger than 21 years of age, prisoners, pregnant or breastfeeding women Patients unable to provide informed consent Any medical condition which makes the candidate an inappropriate subject for study participation, in the investigator's judgment